37 Fraunhofer Press box https://t1p.de/omlnf Share Source: © wbk Engineers and data scientists observe the production with the aid of the learned models and jointly develop data-driven control of the process, which deliberately adapts its behaviour. In this way, the market can be served with new products much earlier. With a considerably shorter time-to-market, target markets for new products can be served before the production processes have been optimised with regard to manufacturing costs, e.g. in the production of battery cells or modules, automated production of electric drives or the production of fuel cells. Systematic procedure through AI-system engineering Especially in industrial production, the use of AI methods and tools must ultimately be systemised so that they ensure reliable engineering which complies with requirements (functional and IT), safety, and robustness. We refer to this as “AI Systems Engineering”. AI Systems Engineering involves the development and operation of AI-based solutions as an integral component of systems which fulfil complex tasks. Therefore, AI Systems Engineering complements basic research on AI and ML and forms a bridge to engineering sciences. The objective is to make AI and ML methods usable by engineers on the basis of typical questions and methods, and to transfer these into suitable system architectures by means of IT methods, e.g. with regard to availability, resilience and extendibility. The Competence Center for AI System Engineering (CC-KING) in Karlsruhe is especially dedicated to these tasks. This also includes AI-related tasks, which form the focus of the Karlsruhe Research Factory. The Karlsruhe Research Factory Every company needs a roadmap for its own route to the digitalisation of its products and processes. A facility for this is the long-term Fig. 1: The Karlsruhe Research Factory for AI-Integrated Production and targeted cooperation by industrial partners with the Karlsruhe Research Factory. On a production area of 5,000 m² with stateof-the-art infrastructure, we implement industrial AI projects together with partners from industry (Fig. 1). The initiators of the Karlsruhe Research Factory already work in an interdisciplinary manner: materials and process engineering experts (Fraunhofer ICT), smart manufacturing and assembly engineering (wbk at the Karlsruhe Institute of Technology - KIT) as well as information technology and industrial AI (Fraunhofer IOSB) form the basis for this. The benefits for industrial partners are the equipment, existing solution modules in which time-consuming development has already been thought through and can be rapidly deployed, as well as the proximity to qualified personnel from cutting-edge research in AI and ML at the Fraunhofer Institute and the KIT. In addition to conventional project cooperation, the partners in the Research Factory offer so-called ‘Corporate Innovation Labs’, in which companies can send their employees as “Embedded Scientists” for a defined period. These work together as a team with Fraunhofer and/or KIT experts on solutions for innovative products, new production processes or data-based services which are defined by the respective company. A steering group reviews the results at regular intervals and if necessary inputs new or modified tasks into the innovation process. The “Embedded Scientists” have the option of achieving a PhD qualification, and on completion return to their companies as multipliers. Outlook With Industry 4.0, Germany has a leading expor t – however the possibilities of Industry 4.0 are by no means exhausted. For us, this means regarding the digitalisation of products, production processes and their equipment, as well as the associated IT systems and infrastructures as integral components, e.g. in relation to the mastery of innovative production processes. The utilisation of data-driven AI and ML methods in industrial applications will result in a strategic advantage over competitors. If Germany and its companies wish to defend their leading position in the production of high-quality goods, they must also take a leading position in this field. The prerequisite for the effective use of artificial intelligence and machine learning is that application-related research in this field is merged with conventional engineering disciplines to bridge the differences in methods and procedures. The Karlsruhe Research Factory is designed for this task and addresses developments together with industrial partners. We are always on the lookout for interesting use cases and “difficult challenges” from industry, which can only be resolved in an interdisciplinary manner. Podcast Oliver Schonschek in an interview with Dr. Olaf Sauer, Embedded Scientists at the Karlsruhe Research Factory https://t1p.de/jzl7
RkJQdWJsaXNoZXIy MjUzMzQ=